This study investigates the relationship of drinking water quality with cancer of the colon in North Carolina using the case control approach. A matched series of living colon cancer cases and hospital controls will be compared for differences in quality of their drinking water and other environmental factors. A substudy will determine whether past diet is a confounding factor. Sample size and methodology have been chosen to detect small differences in risk, and water quality indices have been constructed for periods 20 and 30 years in the past to account for the issues of latency. These indices, which are based on source, treatment and prior use characteristics, provide a basis for establishing a gradient of carcinogenic exposure. With this information, estimates of risk for each exposure level will be obtained and comparisons between these will be made. Using the rigorous approach, it will be possible to provide meaningful data on the importance of water borne carcinogens in the etiology of a leading cancer site.